Constructing a neural network by hand

Neural Networks belong to the most widely used and most popular machine learning algorithms as they can be used for a variety of purposes, such as for example image recognition or predictions. However Neural Networks are often said to be very complex and rather uninterpretable. Here I want to show, by means of a simple example, that the reputation of being complex can partially be disprove, as I will code a neural Network by hand.

DNA sequence classification via an EM algorithm

SeqEM ia a new genotype-calling approach based on the EM algorithm which is supposed to be more precise and accurate than the methods that have existed until now. The goal is the implementation of the so-called SeqEM algorithm in R by hand.

Bootstrap hypothesis testing on quantile value

The bootstrap method and traditional statistics use very different approaches to estimate sampling distributions. Bootstrap hypothesis testing is very popular as it does not assume any underlying distribution of the data (unlike the traditional approach). As there is currently no predefined function in *R* that allows to perform bootstrap hypothesis testing on quantile values we are going to do it by hand.